Bug#1013178: transition: ceres-solver

2022-06-21 Thread Pierre Gruet

Hi Emilio,

On Tue, 21 Jun 2022 09:49:32 +0200 Emilio Pozuelo Monfort 
 wrote:

> [...]
>
> > All reverse dependencies are building fine at least on amd64 [2].
>
> That link doesn't tell me if the rdeps build against the new SONAME. 
Have you

> tested that? If so, go ahead.

Thanks for looking at this transition bug. Yes, I can confirm I 
successfully built the rdeps against the new SONAME a few days ago.


So François will go ahead within the upcoming days.

>
> Cheers,
> Emilio
>
>

Best,

--
Pierre


OpenPGP_signature
Description: OpenPGP digital signature


Processed: Re: Bug#1013178: transition: ceres-solver

2022-06-21 Thread Debian Bug Tracking System
Processing control commands:

> tags -1 confirmed
Bug #1013178 [release.debian.org] transition: ceres-solver
Added tag(s) confirmed.

-- 
1013178: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1013178
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems



Bug#1013178: transition: ceres-solver

2022-06-21 Thread Emilio Pozuelo Monfort

Control: tags -1 confirmed

On 18/06/2022 15:29, Francois Mazen wrote:

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition
X-Debbugs-Cc: franc...@mzf.fr

Dear release team,

I and Pierre Gruet (pgt@d.o) would like to transition ceres-solver to the new
SOVERSION (3).

The upstream changed the SOVERSION of the package without changing the major
version number (2.1.0). That's why we missed this ABI change, and Pierre
reverted the upload by uploading a new package with the +really suffix
(2.1.0+really2.0.0).
Upstream does not follow semantic versioning and confirmed that this behavior
is intentional [1].
So, a transition process is needed for the Debian package to handle this
SOVERSION update.


This transition clashes with the ongoing onetbb through openturns, however those 
two have already migrated to testing so that shouldn't be a blocker.



All reverse dependencies are building fine at least on amd64 [2].


That link doesn't tell me if the rdeps build against the new SONAME. Have you 
tested that? If so, go ahead.


Cheers,
Emilio



Bug#1013178: transition: ceres-solver

2022-06-18 Thread Francois Mazen
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition
X-Debbugs-Cc: franc...@mzf.fr

Dear release team,

I and Pierre Gruet (pgt@d.o) would like to transition ceres-solver to the new
SOVERSION (3).

The upstream changed the SOVERSION of the package without changing the major
version number (2.1.0). That's why we missed this ABI change, and Pierre
reverted the upload by uploading a new package with the +really suffix
(2.1.0+really2.0.0).
Upstream does not follow semantic versioning and confirmed that this behavior
is intentional [1].
So, a transition process is needed for the Debian package to handle this
SOVERSION update.

All reverse dependencies are building fine at least on amd64 [2].

Best Regards,
François

[1] https://github.com/ceres-solver/ceres-solver/issues/824
[2]
https://buildd.debian.org/status/package.php?p=colmap,openturns,sight=compact


Ben file:

title = "ceres-solver";
is_affected = .depends ~ "libceres2" | .depends ~ "libceres3";
is_good = .depends ~ "libceres3";
is_bad = .depends ~ "libceres2";